1 Introduction

Tin dioxide (SnO2) has attracted lots of attentions in wide applications such as the detection of inflammable gases, volatile organic compounds, and toxic gases, due to its unique physical and chemical properties. SnO2 is an n-type semiconductor with tetragonal rutile structure and large energy band gap of 3.6 eV at 300 K [1]. Till now, there are many different methods to prepare SnO2 thin film, for example, sol–gel [2], thermal evaporation [3], chemical vapor deposition (CVD) [4], radio frequency (RF) magnetron sputtering [5], and spray pyrolysis [6]. Among these methods, the sol–gel method has been widely investigated because of its many advantages such as low reaction temperature, low cost, and easy process [1]. However, Imad et al. [7] found that the sol–gel method without adding glycerin lead to produce thin films suffered from the formation of cracks.

Hydrogen (H2) is a kind of more efficient and clean source of energy which has been used in automobiles, aircraft, fuel cells, and chemical industries, etc. [8, 9]. Since H2 gas is colorless, odorless, highly volatile, and inflammable [9], the detection at room temperature (RT) is very important for chemical industries and environmental applications. Detection sensor is a usual method to alarm the formation of potentially explosive mixtures of H2 in air ambient and therefore to prevent the risk of explosions and fires [10–12]. RT H2 gas sensor also attracts much attention in other fields because of their particularly low power consumption [13], the ability to be used safely in flammable environments [14], and long lifetime [15].

Recently, gas sensing mechanism for SnO2 films has been studied for different periods [16]. Efforts have also been done for verifying their RT detection by using nano-sized SnO2 and applying dopants in the SnO2 thin films [17]. Usually, the gas sensing tests were carried out at high operation temperature [18]. Hamaguchi et al. [19] reported that the H2 gas sensor can not be performed at low temperature and therefore to extend the response and recovery time.

This paper focuses on fabrication of functional nanocrystalline SnO2 thin films and performance of H2 gas sensors at RT for different gas concentrations. The main goal of this study is to take the advantage of adding glycerin to the sol solution to solve the cracks problem of nanocrystalline SnO2 thin films on bare Si (100) substrates, and then to fabricate thin film gas sensor that could emerge reasonable sensitivity.

2 Experimental

The p-type (100) silicon wafer (10 mm × 10 mm) was cleaned by the standard Radio Corporation of America (RCA) method. Nanocrystalline SnO2 thin films were grown using sol–gel spin coating method [2, 20]. 0.1 M tin (II) chloride dihydrate (SnCl2·2H2O) was dissolved via 70 mL of pure ethanol (C2H5OH) and placed in covered flasks. The resulted sol solutions placed in closed flasks and stirred by magnetic stirrers for 3 h and kept at 70 °C for 8 and 10 h, respectively. Moreover, glycerin (C3H8O3) was added to a volume ratio of 1:12 in order to eliminate cracks [7]. The process of preparing the sol solutions was separately completed at RT for the remainder of the 24 h. Thereafter, the sol solutions were spin-coated on Si (100) substrates at a rotation speed of 3000 rpm for 30 s. The as-deposited films were oven-dried at 100 °C for 10 min, to obtain high thickness, spin coating and drying operations were repeated 10 times for all samples at different aging heat times. The whole samples were annealed at 400 and 500 °C in air ambient for 2 h in order to achieve the crystallization of SnO2.

The fabrication of the metal–semiconductor–metal (MSM) gas sensing devices was conducted via RF sputtering of Pd grid using a shadow mask on top of nanocrystalline SnO2 thin films. This mask contains two electrodes and each electrode consists of four fingers, the space between two neighboring fingers is 0.4 mm, and the width of each finger is 0.35 mm as shown in Fig. 1.

Fig. 1
figure 1

Schematic of Pd grid contact deposited on the nanocrystalline SnO2 thin films

Full size image

Gas sensing was performed in a homebuilt gas chamber, which was fabricated by an acrylic plastic box and heated by a small ceramic heater that joined to a temperature controller. The chamber was joined to three mass flow meters/controllers: the first for 0.1 % H2/balance N2 gas, the second for the carrier N2 gas, and the third for air ambient that is provided by an air pump via a drying tube containing fine silica gel. These mass flow meters/controllers used to supply a constant flow (1000 sccm) during test measurements. In this case, the resulted concentration of H2 gas (c 2: in ppm) was computed by the following eq. [18, 21] c 2 (ppm) = (c 1 × F H)/F total, where c 1 is the concentration of H2 in the bottle (0.1 % = 1000 ppm), F H is the gas flow of 0.1 % H2 balanced with N2, and F total is the total flow of (0.1 % H2/balance N2) and the flow of diluting N2 gas. The relative sensitivity (S) of the nanocrystalline SnO2 thin films gas sensor can be defined as the relative change in the conductivity (\(\Delta G\)) upon exposure to H2 gas [22, 23]:

$$S\left( \% \right) = \frac{\Delta G}{{G_{\text{air}} }} \times 100 = \frac{{\left( {G_{\text{g}} - G_{\text{air}} } \right)}}{{G_{\text{air}} }} \times 100$$
(1)

where G air is the conductivity in an air ambient and G g is the conductivity in the presence of the gas being sensed.

Also the relative sensitivity (S) could be rewritten in terms of the electric current generated in the semiconductor-based gas sensor on the application of a constant bias voltage [24, 25]:

$$S\left( \% \right) = \frac{{\left( {I_{\text{g}} - I_{\text{air}} } \right)}}{{I_{\text{air}} }} \times 100$$
(2)

where I g is the current measured in the presence of the gas being sensed and I air is the current measured in air ambient.

The crystal structure of the fabricated samples of the SnO2 thin films was performed using X-ray diffraction (XRD) analysis of PANalytical X’ pert Pro MRD equipped with a Cu Kα radiation of (λ = 0.154060 nm). The morphologies were characterized by field emission scanning electron microscopy (FESEM) of model Leo-Supra 50VP, Carl Zeiss, Germany. A current source (2400 Source Meter, Keithley, Cleveland, Ohio, USA) that was joined to a computer through the LabTracer (test integration software) was utilized to measure the electrical current passing in the gas sensing device on the application of a bias voltage.

3 Results and Discussion

Figure 2 shows XRD patterns of nanocrystalline SnO2 thin films grown on bare Si (100) substrates, in which Fig. 2a, b are for films without adding glycerin and annealed at 500 and 400 °C, respectively. The peaks correspond well to standard bulk SnO2 with a tetragonal rutile structure (JCPDS card No. 041-1445) [26, 27]. While the peaks in Fig. 2c, d become sharper and stronger after adding glycerin to sol solutions. This is due to the annealing process which enhanced the crystallization of films, and therefore increased crystallite size, and reduced defects [27, 28].

Fig. 2
figure 2

XRD patterns of nanocrystalline SnO2 thin films annealed at 500 and 400 °C a, b without glycerin, and c, d with glycerin

Full size image

The average crystallite size (D) of nanocrystalline SnO2 thin films was calculated using the (110) major and first diffraction peak by Debye-Sherrer formula [7]:

$$D = \frac{0.9\lambda }{\beta \cos \theta }$$
(3)

where λ = 0.1540 nm, β is the full width at half maximum intensity of the distinctive peak, and θ is the Bragg’s angle. It can be seen that the crystallite sizes increase after addition of glycerin as shown in Table 1.

Table 1 Crystallite size of SnO2 thin films obtained under different conditions
Full size table

FESEM images clearly show the morphology evolution of the SnO2 film grown on Si substrates with and without adding glycerin. Figure 3a, b show images of films without adding glycerin and annealed at 500 and 400 °C, respectively. One can notice the existence of cracks on the film. These cracks can produce negative effect on the performance for any device [7, 29]. Small nanoparticles with irregular size were observed to confirm the polycrystalline structure of the films. However, as shown Fig. 3c, d, there are no any cracks observed for nanocrystalline SnO2 films. This is due to the contribution of glycerin added to the sol solution with a volume ratio of 1:12 [7]. In addition, the particle size increases and becomes more regular and homogeneous. Figure 3c shows the surface morphology under the first aging heat time of 8 h at 500 °C. It is better than the second aging heat time of 10 h at 400 °C (Fig. 3d). This can be understood that increase of annealing temperature could enhance the crystallization degree of the films [27, 28].

Fig. 3
figure 3

FESEM images of nanocrystalline SnO2 thin films annealed at 500 and 400 °C a, b without glycerin, and c, d with glycerin

Full size image

Figure 4 shows the I-V characteristics of nanocrystalline SnO2 films sensors, in which Fig. 4a, b show Ohmic behaviors, due to allow for contact between electrodes and bare Si substrates, whereas Fig. 4c, d show Schottky contact as a result of the work function for Pd element is higher than for SnO2 [30, 31].

Fig. 4
figure 4

The I-V characteristics of nanocrystalline SnO2 thin films annealed, respectively, at 500 and 400 °C (a, b) without glycerin, (c, d) with glycerin, respectively

Full size image

The sensitivity (S) of the sensors with a Pd–nanocrystalline SnO2–Pd (MSM) configuration is denoted by Eq. (1). For Schottky behaviors, the sensitivity emerged at different temperatures (RT, 75, and 125 °C) and at different H2 gas concentrations, while there is no sensitivity on the sensors with Ohmic behaviors. The sensitivity of the sensors measured under different conditions are shown Fig. 5 (Device 1: 8 h at 500 °C) and Fig. 6 (Device 2: 10 h at 400 °C), where the exposure pulses is 1000 ppm H2/balance N2 and dry air, as well as the bias voltage is 0.2 V and 1 V, respectively. From Fig. 5a, the RT sensitivity of nanocrystalline SnO2 thin film sensor was found to be 120 %. It drifts from zero level during the repeated cycling of H2/balance N2 and dry air. This may be due to incomplete removal of H2 gas on the surface of nanocrystalline SnO2 films and the inefficient adsorption of O2 during the dry air injection into the gas chamber [32]. In addition, the sensitivity and repeatability increase with increasing the operating temperature are shown in Fig. 5b, c, respectively. The increases may be attributed to the increase of the adsorption/desorption of gas molecules in presence of different oxygen species [33].

Fig. 5
figure 5

The sensitivity and repeatability of gas sensor based on nanocrystalline SnO2 thin films upon exposure to successive pulses of 1000 ppm H2/N2 gas and dry air at different sensing temperatures a RT, b 75 °C, and c 125 °C. The films synthesized with adding glycerin, aging 8 h, and annealing at 500 °C

Full size image
Fig. 6
figure 6

The sensitivity and repeatability of gas sensor based on nanocrystalline SnO2 thin films upon exposure to successive pulses of 1000 ppm H2/N2 gas and dry air at different sensing temperatures a RT, b 75 °C, and c 125 °C. The films synthesized with adding glycerin, aging 10 h, and annealing at 400 °C

Full size image

It is clearly observed that the sensitivity of Device 1 outperforms better than the latter as shown in Fig. 6, where the sensitivity value was 95 % at RT. This variation in the sensitivity value is due to the improvement of crystallization, which allowed for increasing in the electron–hole transport, and therefore the response to H2 increases [34]. The computed power consumption of two SnO2 films sensors is respective 65 and 86 µW for the detection H2 at RT. This makes that these sensors are easy to use in remote area where the available power may be limited. A significant variation in the sensitivity value of nanocrystalline SnO2 thin film sensors was observed upon exposure to H2 gas concentrations from 150 to 1000 ppm at different operating temperatures (shown in Figs. 7 and 8). Furthermore, one can see in Fig. 7a, b the sensitivity drift from the baseline. This is due to both incomplete recovery of H2 gas sensors and inefficient adsorption/desorption of the gases when the device is operated at low temperature compared with the operation of the H2 gas sensor at high temperatures (see Fig. 7c) [33]. Figure 8 displays the sensitivity value for Device 2 which is less than that of Device 1 at different H2 gas concentrations, which agree to the results in Fig. 6. The response time of H2 sensing is denoted as the time required to the final saturation state current was 90 % from its top value, while the recovery time is denoted as the time required to 10 % value of the saturation state current [33]. Table 2 displays the response and recovery times of nanocrystalline SnO2 thin film (Device 1) for the H2 gas sensing and their relative sensitivities at different operating temperatures were then compared with the previous studies.

Fig. 7
figure 7

The sensitivity of gas sensor based on nanocrystalline SnO2 thin films under various H2 gas concentrations (150–1000 ppm) at different sensing temperatures a RT, b 75 °C, and c 125 °C. The films synthesized with adding glycerin, aging 8 h and annealing at 500 °C

Full size image
Fig. 8
figure 8

The sensitivity of gas sensor based on nanocrystalline SnO2 thin films under various H2 gas concentrations (150–1000 ppm) at different sensing temperatures a RT, b 75 °C, and c 125 °C. The films synthesized with adding glycerin, aging 10 h, and annealing at 400 °C

Full size image
Table 2 Performance comparison of nanocrystalline SnO2 thin film sensor (Device 1) with previous studies
Full size table

Figures 5, 6, 7 and 8 show a clear increase of sensitivity with high stability of nanocrystalline SnO2 thin film sensors as the operating temperature increases. This behavior is because of an enhancement in the adsorption/desorption processes of gas molecules, where there exist different types of oxygen species at higher operating temperatures [33]. Thereby, the reaction probability between oxygen species and H2 gas will be increased. Thus, the removal of H2 gas from the surface of nanocrystalline SnO2 thin film sensors is enhanced. As a result, it could be observed that the sensitivity is not drifted from the baseline.

The sensing mechanism of nanocrystalline SnO2 thin films is related to the reduction of the exposure to gas like H2 by the adsorption of oxygen molecules on nanocrystalline SnO2 surface. When the surface of nanocrystalline SnO2 thin films is exposed to air ambient, the oxygen species will be adsorbed. Depending on the operating temperatures, the adsorbed oxygen species will capture electrons from nanocrystalline SnO2 thin films surface and become negatively charged, which will increase the depletion region and therefore increase the resistivity [35]. There are different oxygen species depending on the operating temperatures, which can be described as follows: [36].

$${\text{O}}_{ 2} \left( {\text{gas}} \right) \to {\text{O}}_{ 2} ({\text{ads}})$$
(4)
$${\text{O}}_{ 2} \left( {\text{ads}} \right){\text{ + e}}^{ - } \to {\text{O}}_{ 2}^{ - } \left( {T < 100^\circ {\text{C}}} \right)$$
(5)
$${\text{O}}_{ 2}^{ - } \left( {\text{ads}} \right){\text{ + e}}^{ - } \to 2{\text{O}}^{ - } (T = 100 - 300^\circ {\text{C}}).$$
(6)

H2 molecules are dissociated to H atom on the Pd contact, which diffused to the surface of nanocrystalline SnO2 thin films, and reacts very quickly with different adsorbed oxygen species by negative charges [33]. Thereby the electrons captured by the oxygen species will return back to the conduction band of thin film, resulting in an increase of electron concentration in the conduction band so that the resistance of nanocrystalline SnO2 will reduce. The sensing reactions could be explained by using the following chemical reactions: [37, 38].

$$2{\text{H}}_{2} + {\text{O}}_{2}^{ - } ({\text{ads}}) \to 2{\text{H}}_{ 2} {\text{O}} + {\text{e}}^{ - }$$
(7)
$$2{\text{H}}_{2} + {\text{O}}^{ - } ({\text{ads}}) \to 2{\text{H}}_{ 2} {\text{O}} + {\text{e}}^{ - } \left( {T = 100 - 300^\circ {\text{C}}} \right)$$
(8)
$$4{\text{H}} + {\text{O}}_{2}^{ - } ({\text{ads}}) \to 2{\text{H}}_{ 2} {\text{O}} + {\text{e}}^{ - } .$$
(9)

The sensitivity of H2 gas sensor will reduce when the nanocrystalline SnO2 thin films are exposed to air ambient again, where the air ambient inputs to the gas chamber containing oxygen species. Thereafter, the air oxygen will react with the chemisorbed H2 on the surface of nanocrystalline SnO2 thin films. Hence, the resistance of the nanocrystalline SnO2 thin films goes back to its initial value [33].

4 Conclusions

Nanocrystalline SnO2 thin films were grown on bare Si (100) substrates using a simple cost-effective sol–gel method. The cracks that appeared on the surface of thin films were avoided by adding glycerin to the sol solutions in volume ratio of 1:12. Two devices with Schottky contact were fabricated to detect H2 gas with different concentrations and different temperatures. One of the devices shows a stable sensitivity of 120 % at RT with power consumption of 65 µW, which is appropriate in remote regions. The good sensitivity is attributed to the high porosity of nanocrystalline SnO2 thin film generated by adding glycerin. It makes easy for the adsorption/desorption of gas molecule. Moreover, Pd finger contacts significantly enhance the sensing properties of the gas sensor. The results also show that the good crystallinity of thin film can enhance the performance of device.